Yale ICF Working Paper No.

09-12
June 2009

Hedging Real-Estate Risk

Frank J. Fabozzi
Yale School of Management

Robert J. Shiller
Yale School of Management

Radu Tunaru
Cass Business School
City University, London

Electronic copy available at: http://ssrn.com/abstract=1419528

 Hedging Real-Estate Risk
Frank J. Fabozzi, Robert J. Shiller, and Radu Tunaru

Frank J. Fabozzi is Professor in the Practice of Finance, Yale School of Management,

New Haven, CT
frank.fabozzi@yale.edu

Robert J. Shiller is Arthur M. Okun Professor of Economics, Yale University,

New Haven, CT
robert.shiller@yale.edu

Radu Tunaru is Senior Lecturer in Financial Mathematics, Cass Business School, City
University, London
r.tunaru@city.ac.uk

Electronic copy available at: http://ssrn.com/abstract=1419528

Hedging Real-Estate Risk

Real-estate assets represent more than one-third of the value of all the underlying physical

capital in the United States and the world. The relationship between the level of interest rates

and housing prices does not always follow one direction and a shock event in one market may

trigger deep repercussions in the other. With the spread of the securitization process, the risks

rooted in these two fundamental markets can have far reaching outcomes.

Prices in the residential housing market are determined by direct trade between buyers

and sellers who are influenced by emotional involvement and other opaque social factors such

as change of employment or change of school for children. Residential real-estate assets are

naturally not diversified and are a combination of a consumption asset and a leveraged

investment. As Shiller and Weiss [1999] point out, this characteristic poses greater risk for the

financial stability of individuals due to geographic fluctuations in property prices. Ideally,

homeowners could use derivatives on relevant real-estate indices to hedge this risk and

stabilize prices. Although real-estate derivatives should be preferred to insurance-type

contracts because of direct settlement, liquidity is very important and this can be established

only after banks decide to participate more actively in the real-estate index futures and options

markets, as advocated by Case, Shiller, and Weiss [1993]. Housing prices are sticky when they

are going down, sellers being reluctant to sell at a price below the price at which the property

was initially purchased. Hence, the market is localized, the information is asymmetric, and

participants’ price expectations are very much influenced by the recent series of prices. 1 While

it would be difficult for homeowners to hedge directly the price of their homes, banks and

building societies that have mortgage portfolios more diversified nationally should be enticed

to hedge their exposure with derivatives written on local indices. The use of index-based

futures contracts and options for hedging mortgage risk, default risk, and real-estate price risk

1
This is discussed in more detail in Case, Quigley, and Shiller [2003]

Electronic copy available at: http://ssrn.com/abstract=1419528

has long been advocated by Case and Shiller [1996]. Fisher [2005] provides an overview of

NCREIF-based swap products and Clayton [2007] examines various indices developed for

derivatives trading

The introduction of derivatives in the real-estate market is not easy because liquidity is

difficult to establish when returns are predictable. An extensive discussion highlighting the

important psychological barriers that need to be passed for the establishment of real-estate

derivatives is provided in Shiller [2008]. Carlton [1984] argues that if changes in market prices

are predictable, then changes in prices cannot be perceived as risky. The major obstacle for the

introduction of real-estate derivatives was that when returns follow trends at certain points in

time, then market sentiment is in only one direction and it is difficult to find counterparties

trading against the trend. Nevertheless, if a futures contract is already trading for a series of

future maturities, then the shape of the forward curve on real-estate index becomes important.

Trades may be executed on the curve, say short a futures with a long maturity and long a

futures with shorter maturity, which would be impossible to execute otherwise. With futures

and options on futures, an entire spectrum of trading strategies becomes available and market

participants such as hedge funds, investment houses, and private equity funds may provide

much needed liquidity. 2

The exponential growth of the subprime mortgage market from 2002 to 2007 was

driven by the exploitation of securitization as a process of ring-fencing the risks of a collateral

portfolio on one side and the introduction of the real-estate collateralized debt obligation (RE

CDO) concept on the other side. The ever-growing demand for credit risky bonds pushed the

boundaries of this new market into new territory, residential mortgage-backed securities

(RMBS) backed by home equity loans (i.e., loans to credit impaired borrowers) and

2
For example, Alpha Beta Fund Management (ABFM) announced in July 2007 that it targeted pension
funds and other investors that may benefit from access to British housing by dealing in an over-the-counter
property derivatives market which tracked the Halifax House Price Index (HHPI)

3
commercial mortgage-backed securities (CMBS). The cash flows of a RMBS, CMBS, and RE

CDO depend fundamentally on the performance of a pool of mortgage loans, which in turn

depend on the behavior of individual homeowners and commercial borrowers. These real-

estate structured products are radically different from a CDO where the collateral consists of

corporate credits whose corporate names could be monitored and their balance sheet

scrutinized regularly. Hence, the real-estate risk drivers ─ prepayments and defaults clustering

and timing ─ as well as recovery rates, could influence the financial stability of companies and

institutions not directly related to the spot real-estate market. Hedging the potential disruption

of scheduled cash flows is not an easy task and there are only a few instruments available in

the market.

The advances in futures markets on real-estate indices may improve efficiency in spot

markets and improve price discovery. Since transaction costs are high and create a barrier for

entry into spot markets, futures markets may also help indicate the level of spot prices in the

future and current market volatility. Another benefit of real-estate derivatives is that they are

useful tools that allow investors access to an important asset class that would be hard to access

otherwise. Furthermore, due to the lack of correlation of housing prices with equity prices,

expanding diversified portfolios to include real estate could be highly beneficial, particularly

for insurance and pension-funds. 3 Englund, Hwang, and Quigley [2002] point out that there

could be large potential gains from instruments that would allow property holders to hedge

their lumpy investments in housing.

Obviously, the first step in hedging is the selection of a suitable hedging instrument. A

primary factor in deciding which derivative contract will provide the best hedge is the degree

of correlation between the factors that drive the price of the derivative instrument under

consideration as the hedging vehicle and the underlying risk that investors seeks to eliminate.

3
This is more extensively discussed in Webb, Curcio, and Rubens [1988] and Seiler, Webb, and Myer
[1999]

4
Correlation is not, however, the only consideration when the hedging program is of significant

size. If, for example, an investor wants to hedge a very large cash position, liquidity becomes

an important consideration and it might be necessary to split the hedge among two or more

different derivatives.

Real-estate derivatives are useful to several categories of end users. The first category

consists of individuals who are property owners and private investors specializing in real

estate. Although this category of users is very large, in practice, not many may employ

property derivatives due to knowledge and transaction costs barriers. The second category

consists of portfolio managers hedging their price risk exposure in both domestic and foreign

real estate. An adjacent category contains dealers and portfolio managers in structured

products seeking to hedge their positions. Finally, real-estate derivatives can be embedded by

structurers into newly designed structured products. The risks that users in these categories are

hedging with property derivatives may vary. For example, while the members of the first

category will hedge price risk, the users in the other categories may also consider property

derivatives in connection with interest-rate risk and, possibly, currency risk.

In this paper, we describe the real-estate derivatives available worldwide and discuss

the issues related to the pricing of these instruments and to the managing of hedging

instruments over time.

REVIEW OF INSTRUMENTS USED FOR HEDGING REAL-ESTATE RISK

In this section, we describe the development of real-estate linked derivatives worldwide. The

most developed markets from a financial product innovation point of view are in the United

States and the United Kingdom, although some activity has been noticed in other developed

European countries as well. The instruments can be classified by the type of real-estate risk

5
they are hedging (1) housing price risk, (2) commercial property price risk, and (3) mortgage

loan portfolio amortizing risk.

Hedging Housing Price Risk

A major component of the real-estate asset class is represented by residential housing.

Housing prices are determined by macroeconomic conditions and by the behavior of the

individuals buying and selling properties. Housing price risk is mainly associated with the

sharp downturn or fall in housing prices. In addition to the owner of the property, this risk is a

major concern to banks and other lending institutions and investors in structured products

backed by residential mortgage loans. Here we review the financial instruments that have been

designed to be used for hedging this risk.

In May 2006, the Chicago Mercantile Exchange (CME) launched futures and options

on futures trading on the Standard & Poor’s/Case-Shiller Home Price Index, an index

constructed based on repeated sales analysis. As of 2009, there are futures contracts with

maturities extending out 18 months into the future, listed on a quarterly cycle of February,

May, August, and November; futures contracts with maturities extending out 19 to 36 months

into the future, listed on a bi-annual schedule May and November; and futures contracts with

maturities extending out 37 months to 60 months into the future, listed on an annual schedule

with November maturity. The futures contracts trade at $250 times the index with a tick of

$50, while the options trade on one futures contract with a tick of $10, for a range of strikes at

five index point intervals from the previous day close price of the futures on the Case-Shiller

Index. There are futures for 10 U.S. cities and also an aggregate index

The RPX is a residential index developed by Radar Logic Incorporated 4 that captures

owner-occupied housing in 25 U.S. metropolitan statistical areas (MSAs). There is a global

4
More details and price fixings can be found at www.radarlogic.com

6
MSA 25 Composite index that reflects the top 25 MSAs, as well as indices for the individual

MSAs. This index is based on rolling quarterly price fixings, measured by price per square

foot, 5 and updated daily. Trading on derivatives contingent on the RPX index started in

September 2007. The first RPX index-based derivative traded was a total-return swap. For this

derivative, a fixed payment is periodically exchanged for the growth of the RPX index.

Subsequently, RPX forward contracts began trading in May 2008 and they have become the

most liquid contracts based on the RPX index.

In June 2009 the firm MacroMarkets LLC, has scheduled a launch of securities whose

underlying value is linked by formula to the S&P/Case-Shiller Composite 10 Home Price

index, an index of home prices in major metropolitan areas, on the New York Stock Exchange.

There are two five-year securities, with ticker symbols UMM (for Up Major Metro) and DMM

(for Down Major Metro). The securities are automatically created or redeemed in pairs by

authorized participants, and the funds contributed at creation are invested in US Treasury Bills.

Thus the pair together is fully collateralized and represents a balanced book for the issuer, so

there is no counterparty risk. The elements of the pair trade separately, so there is price

discovery for real estate five years after issue.

In the United Kingdom, real-estate derivatives are written on a house price index (HPI),

the most common being the Halifax (HHPI) series. The HHPI is the longest running monthly

housing price series in the United Kingdom with data available since January 1983. This index

is based on the largest monthly sample of mortgage data, typically covering around 15,000

house purchases per month. This is a hedonic index.

In 2003, Goldman Sachs issued the first series of a range of covered warrants based on

the Halifax All-Houses All-Buyers seasonally-adjusted index on the London Stock Exchange

(LSE). More recently, based on the HPPI, in August 2007, Morgan Stanley agreed on an exotic

5
This index can also be used to account for values paid in arms-length residential real estate transactions on
a price per square foot basis.

7
swap with an undisclosed counterparty, worth more than £1 million. This is the UK’s first

residential property derivative trade that included an embedded exotic option, a “knock-in put”

option. This derivative allows the counterparty to gain if the HHPI rises, subject to a maximum

payout. The investor’s capital is protected unless the HHPI falls below an initially specified

value.

Hedging Commercial Property Price Risk

The other major component of real-estate asset class is commercial properties. Price

risk for commercial properties is similar to housing price risk but it is not generally influenced

by the behavior of the same participants as in the residential housing market. Commercial

property prices are determined by supply and demand and specialized market participants.

While some degree of correlation is expected between commercial price risk and housing price

risk, there are sufficient differences to warrant a separate analysis and separate hedging

instruments 6 .

A major property index used as an underlying for property derivatives to hedge

commercial property price risk in the United States is the NCREIF series and, in particular, the

National Property Index. The NCREIF property index is based on an aggregation of appraised

property values. This index is the underlying for two types of swaps that were traded to date.

One is the total-return swap that allows an investor to synthetically reproduce the economic

gains of the index return. The other is an instrument used to swap different NCREIF property

sectors. In June 2008, the NCREIF Property Index Total Return Swaps changed from a

quarterly to annually payment schedule, all contracts paying at the end of the fourth quarter.

6
Although discontinued in early 2009, the first commercial property derivative was based on the Standard
& Poor’s and Global Real Analytics/Charles Schwab Investment Management developed the S&P/GRA
Commercial Real Estate Index, SPCREX™. This family of indices comprised 10 commercial real estate
indices: a national composite, five major U. S. regions (Northeast, Mid-Atlantic South, Midwest, Desert
Mountain, and Pacific West), and four property sectors (apartments, office, retail, and warehouse). The
indices were calculated monthly using a three-month moving average and published with a three month lag.

8
In addition, since June 2008, there are two new total-return swap indices traded, the four-year

maturity NCREIF National and five-year NCREIF National.

Barclays Capital began issuing its Property Index Certificates in 1994 and Property

Index-Forwards in 1996, with the index comprising UK commercial property only. The

contracts are currently traded in the over-the-counter (OTC) market and are written on two

monthly indices published by the Investment Property Database (IPD): the IPD Total-Return

and IPD Capital-Growth indices. These contracts have maturities of three to four years.

The property derivatives market has expanded in the United Kingdom from £850

million in 2005, £3.9 billion in 2006 to £7.2 billion in 2007. The IPD Index Property

Derivatives volume in 2008 was £7.73 billion, not far from £8.30 billion in 2007’s record year.

In Europe, in 2007 the trading volume on the IPD French All Office Index 7 was £787 million,

compared to £283 million on the German All Property Index. 8 Other markets where property

derivatives have been traded include Switzerland, Canada, Japan, Italy, Hong Kong, and

Australia. Thirteen major investment banks acquired licenses for the IPD index family. In

Europe, property derivatives written on IPD indices traded OTC with a notional over £18.8

billion in trades executed by October 2008.

Eurex began trading property futures on February 9, 2009. These futures contracts are

annual contracts based on the total returns of the IPD UK Annual All Property Index 9 for

individual calendar years. The futures contract aims to eliminate counterparty risk, to improve

liquidity to the commercial property sector of the real-estate property market, and to attract a

complete range of potential participants in this asset class. Further futures contracts are going

7
The first French property swap on the IPD France Offices Annual Index was traded in December 2006 by
Merrill Lynch and AXA Real Estate Investment Managers.
8
The first option on an IPD index outside the United Kingdom was referenced to the German IPD/ DIX
Index and it was traded in January 2007 with Goldman Sachs acting as a broker.
9
At the end of 2007, the IPD index covered 12,234 properties with a total value of £184 billion - equivalent
to 49% of the UK investment market. The IPD UK Annual Property Index measures unleveraged total
returns to direct UK property investments and it is calculated using a time-weighted methodology with
returns computed monthly and thereafter compounded for the purposes of the annual index construction.

9
to be launched by Eurex on IPD property indexes, such as UK sector indexes (Offices, Retail,

Industrial) and other European indexes (initially France and Germany) on a demand-led basis.

The Australian Stock Exchange introduced its ASX Property-Trust futures based on the

S&P/ASX200 Listed Property-Trust index in 2002. This index is comprised of the 200 largest

listed property investments by market capitalization and most liquid securities in Australia.

Three years later, the Sydney Futures Exchange listed futures based on the Dow Jones

Australia Listed Property-Trust index.

Proposal for improved commercial real estate indices and instruments based on them will

find their way into the market. For example, Horrigan et.al [2009] demonstrate how to construct

segment-specific indices of property market returns from real-estate investment trust (REIT) return

data, bond data, and property holding data and use those indices to make pure, targeted investments

in the commercial real estate market while retaining the liquidity benefits of the well-developed

public market in REITs. It is possible to re-construct the indices at the daily frequency without

significant increases in noise and at various levels of segment granularity. Moreover, it seems that

these new indices suggested by the authors lead transactions-based direct property market indices

during market turns. Thus, these REIT-based commercial property return indices have the potential

to be used to develop hedging strategies in the real-estate market and support derivatives trading.

Mortgage loan portfolio amortizing risk

At the portfolio level, mortgage loans carry two intrinsic risks that require hedging

tools: default risk and prepayment risk. Default risk is the risk of loss of principal and/or

interest due to the failure of the borrower to satisfy the terms of the lending agreement. This

risk is high for RMBS that are backed by mortgage pools containing subprime mortgages.

A prepayment is the amount of principal repayment that is in excess of the regularly

scheduled repayment due. A prepayment can be for the entire amount of the remaining

principal balance (i.e., complete payoff of the loan) or for only part of the outstanding

10
mortgage balance (referred to as a curtailment). Prepayment risk is greater for RMBS than

CMBS because commercial loans have provisions to mitigate prepayment risk (e.g., lockout

periods, prepayment penalties, yield maintenance, and defeasance).

From an investor’s perspective, prepayment risk is the risk that borrowers will prepay

their loan (in whole or in part) when interest rates decline. This action by borrowers would

force investors to reinvest at lower interest rates. Note that if borrowers prepay when interest

rates rise, prepayments are beneficial to investors because proceeds received can be reinvested

at a higher interest rate. From the perspective of a portfolio manager or risk manager who is

seeking to hedge an RMBS position against interest rate risk, prepayment risk exists even if

prepayments occur when interest rates rise. This is because in establishing a hedge, the amount

to be hedged will vary depending on a projected prepayment rate. At the outset of a hedge, an

amortization schedule is projected based on the projected prepayment rate. Actual prepayment

experience of a hedged asset or can cause a deviation between the projected principal

outstanding based on the amortization schedule designed at the outset for the hedging

instrument and the actual principal. This can result in over or under hedging a position. This

stochastic nature of the amortization scheduled due to stochastic prepayments that hedgers face

might more aptly be described as “amortization risk”.

Default and prepayment risks called for special hedging instruments. The rapid growth

of the U.S. subprime mortgage market led to the introduction of home equity credit default

benchmark indices, referred to as ABX.HE indices that started trading in January 2006. 10 One

of the main roles of the ABX.HE indices is to discover the market view on the risk of the

underlying subprime loans. This is a synthetic instrument that allows investors to identify

macro hedges on the subprime sector of the residential housing market. In addition to

10
The trading is offered by CDSIndexCo, a consortium of credit derivative desks. The members contribute
to the ABX.HE indices, which are managed by Markit Group.

11
managing risk, CDO collateral managers use the index to take advantage of any temporary

pricing discrepancies.

The ABX.HE indices consist of five separate sub-indices, one for each of the rating

categories AAA, AA, A, BBB, and BBB–. 11 Each sub-index consists of 20 tranches (of the

same rating as the rating category for that particular sub-index) from 20 home equity (i.e.,

subprime) deals, with each deal represented once in each sub-index. A new set of ABX.HE

indices is launched every six months on January 19 and July 19, referred to as “Roll Dates.”

Closing mid-market prices are published daily for each set of ABX.HE indices. The

administrator for the indices, Markit, employs a filtering process similar to that used by the

British Banker’s Association to calculate LIBOR. This entails taking the quotes received,

discarding those in the top and bottom quartiles, then calculating an arithmetic mean of the

remainder. 12 The key innovation in this product design is the “pay as you go” (PAUG) CDS to

deal with the unique issues associated with credit events when dealing with CDS backed by

residential real-estate loans.

Turning to commercial properties, for CMBS there is a synthetic index that can be used

for hedging. CMBX is a synthetic family of indices, each index referencing a basket of 25 of

the most recently issued CMBS tranches, sorted by rating class. It is a similar type of index to

corporate CDS indices such as the CDX. Hence, the CMBX indices are rolled into a new "on-

the-run" series twice a year with changes in the reference portfolio reflecting the current

CMBS market. The payments follow the PAUG template. With the CMBX index investors can

get synthetic risk exposure to a portfolio of CMBS. There have been five index series issued in

11
Appropriately, the names of the five sub-indices are ABX.HE.AAA, ABX.HE.AA, ABX.HE.A,
ABX.HE.BBB, and ABX.HE.BBB–.
12
To calculate the official fixing value for a particular sub-index, the administrator must receive closing
mid-market prices from the greater of (1) 50% of ABX.HE contributors, or (2) five ABX.HE contributors.
If, on any date, the administrator receives fewer closing prices for a sub-index than the minimum fixing
number, no fixing number is published for that date.

12
the market, 13 each index with seven tranches containing bonds rated AAA, AJ 14 , AA, A,

BBB, BBB-, and BB, respectively. To be included in the CMBX index, a reference CMBS

must have a minimum size of $700 million, secured by at least 50 separate mortgages that are

obligations of at least 10 unaffiliated borrowers, the underlying mortgages must have no more

than 40% of the properties in the same state, and no more than 60% of the properties can be of

the same property type.

With respect to interest rate risk, dealers holding inventory for whole loans waiting to

be securitized or securitized products held in inventory, as well as asset managers, are

exposed to interest rate risk linked to the amortization of mortgage loan portfolio dynamics.

The amortization is stochastic and is driven mainly by prepayment and default speeds. As a

response to this type of hedging problem, a set of financial instruments linked indirectly to

real-estate risk drivers appeared in the market: structured swaps. These swaps have been used

by building societies in the United Kingdom and ABS desks in major investment banks to

hedge interest rate exposure on mortgage loan portfolios.

There are three types of structured swaps: (1) a balance guaranteed swap, (2) a cross-

currency balance guaranteed swap, and (3) a balance guaranteed LIBOR-base rate swap. In a

balance guaranteed swap, the collected coupons on a collateral portfolio of mortgage loans

are exchanged for a reference LIBOR plus a spread. The notional is the total balance of

surviving loans for the period. It is a complex swap because the notional is stochastic, being

determined by the amount of prepayments, defaults, and arrears in the reference portfolio.

The total coupon paid on an exchange date is also stochastic and, even if the notional was

known, because of the different possible mixture of loans surviving in the reference portfolio,

the coupon could differ for the same notional. The typical structure for a balance guaranteed

13
The CMBX Series 6 originally planned for October 25, 2008, was postponed because of the lack of new
issuance since the CMBX 5 launch. This issue will be revisited on April 27, 2009.
14
The AJ tranche is the most subordinate of the AAA rated tranches. It has been added to CMBX in order
to help institutional investors seeking exposure to an additional credit class.

13
swap 15 is shown in Exhibit 1. This swap became part of the securitization process because it

use was mandated by rating agencies in order to hedge the interest rate risk mismatch

between fixed coupons collectable from mortgage loans and floating coupon payments that

have to be paid post securitization to investors. (Given the large volume of RMBS deals

between 2005 and 2007, one can envisage that there are many institutions holding many

positions in this type of swap.) A balanced guaranteed swap is not an instrument to hedge

prepayment risk or default risk because it does not guarantee the balance of the reference

pool; on the contrary, it is exposed to these two risks. The underwriter of a balance

guaranteed swap is exposed to the amortization speed on the reference mortgage loan

portfolio.

INSERT EXHIBIT 1 HERE

A cross-currency balance guaranteed swap is a more complex swap product that deals

with cross-currency deals whereby the coupons on one leg of the swap and also the notional

are determined in a foreign currency. This swap deals with an extra level of risk through its

foreign-exchange exposure. All characteristics described above for a balance guarantee swap

still apply but the notional is stochastic and in a different currency. One subtlety with this

product is that it also has embedded some macroeconomic risk of the country where the

obligors reside. This risk may resurface even if there are no changes in the currencies

specified in the swap. For example, the ebbs and flows of the political and social

environment in the borrowers’ country may cause job losses or price inflation that may

trigger in turn high default rates in the collateral portfolio.

15
There are many problems related to the design of the balance guaranteed swap. One problem is that the
collateral coupon leg is paid in arrears since mortgage payments are collected every day in the period. The
tenor may also differ from deal to deal, the monthly one being more common because this is also the
frequency of mortgage payments and consequently the calculation of defaults and prepayments. Quarterly
is also used in the market on occasion. Since the usual reference floating rate is three-month LIBOR, there
is a basis between three-month LIBOR collected monthly from the swap and three-month LIBOR paid
quarterly to the investors. When the reference rate is one-month LIBOR there would be a basis too.

14
 To understand our next swap ─ a balance guaranteed LIBOR-base rate swap ─ we

need to review the linkages between the official base rate and the standard variable rate

charged on mortgage accounts by lending banks. The official base rate is the rate at which

the Bank of England lends to other financial institutions. The Bank of England’s Monetary

Policy Committee meets every month to determine what needs to be done in response to

economic conditions. While the official base rate plays a very important role in these

markets, banks have the freedom to use a different rate on their loans, which is the standard

variable rate. The standard variable rate is linked directly to the official bank rate, but it is

usually a little higher, reflecting factors such as the real rate at which they borrow from each

other, the business costs associated with lending operations, the volume and maturity profile

of loans, and the funding arrangements. The standard variable rate is likely to vary from bank

to bank and is used to determine the cash flows linked to RMBS. In the United States there

are variable rate mortgages that lend themselves to hedging in the same way.

While a balance guaranteed swap is useful for converting fixed-rate coupons into

LIBOR-based coupons, there are situations when the collateral pool for RMBS and CMBS,

will contain loans that pay variable coupons determined by the variable rate established by

the lender. The level of the variable rate is determined by the lending bank in relation to both

the base rate determined by the national or federal banks and other funding costs driven by

LIBOR-swap rates. In general, the basis between the standard variable rate (or its proxy the

official base rate), and LIBOR is almost constant and not very large during periods of low

volatility of interest rates. However, in turbulent periods, such as the subprime crisis of 2008,

the basis between the standard variable rate and LIBOR may increase dramatically. Mortgage

traders therefore employ our third swap, a balance guaranteed LIBOR-base rate swap, to

hedge this basis risk. One leg of the swap pays LIBOR plus a spread while the other leg pays

the average official base rate during the period. For this type of swap, the LIBOR leg pays

15
the coupon fixed at the beginning of the period but the official base rate leg payment is fixed

at the end of the period. For practical market purposes, the notional can be any value within a

pre-specified band. The buyer has the right to choose the size of the notional at the beginning

of each period. This is a difficult economic decision as only the LIBOR coupon is known.

PRICING AND HEDGING ISSUES RELATED TO REAL ESTATE DERIVATIVES

Derivatives require homogeneity of the underlying for establishing liquidity in their trading.

The lack of homogeneity in real-estate markets was one of the main obstacles to the

development of property derivatives. The securitization mechanism, on the other hand, has

brought a wider participation in these markets and an increased demand for hedging tools.

Case and Shiller [1989, 1990] point out that the housing market in the United States is

inefficient due to serial correlation and inertia in housing prices, as well as in the excess

returns. A possible explanation has been offered by Case, Quigley, and Shiller [2004] who

argue that the price expectation of the majority of market participants is backward looking.

The real-estate literature suggests that repeat-sales indices have three main characteristics:

(1) they are not subject to the noise caused by a change in the mix of sales, (2) they are

highly autocorrelated, and (3) they are predictable with a forecast R-squared roughly 50% at

a one-year horizon. The hedonic indices are subject to model risk stemming from the

multivariate regressions used to build those indices. The factors employed in hedonic

regressions are those characteristics of properties that have been historically found to explain

housing prices. The methods associated with hedonic indices inherit all the common

problems known for regression analysis, namely spurious regression, multi-collinearity, and

model selection.

When pricing or hedging various instruments, it should be remembered that short

sales of real-estate properties is impossible and trading may not be feasible for fractional

16
units. These unique characteristics of real-estate properties have profound implications

regarding valuation principles. Applying the Black-Scholes framework to a market where the

underlying assets have these characteristics would be inappropriate. Moreover, the

predictability in housing prices makes it difficult to establish hedging procedures. However,

one may argue that a well-established futures market may feed information into current

prices and some balance can be achieved with a dual information transmission mechanism.

Given the characteristics of real-estate markets, the valuation of property derivatives

is not straightforward. Neither is the pricing of credit-type instruments such as single-name

ABS credit default swaps (ABCDS), ABS credit default swap indices (ABX.HE), or ABS

CDOs, nor the new cash-flow driven hedging swaps such as balance guaranteed swaps or

LIBOR-base rate swap discussed earlier.

The finance literature in this area is quite sparse and the models proposed are

dichotomized into no-arbitrage models and equilibrium models. 16 No-arbitrage models focus

on the relationship between a real-estate variable and an interest rate.17 They also allow the

calculation of counterparty risk, a very important consideration after the subprime mortgage

crisis. Under the no-arbitrage framework, it can be demonstrated that the spread over a market

interest rate such as LIBOR that a total-return swap payer must charge is highly dependent on the

volatility of the reference index returns and on counterparty risk, with higher volatility of returns

and counterparty risk implying a higher spread over market interest rate. 18

The no-arbitrage argument is difficult to apply because of the impossibility of short

selling and the non-homogeneity of the property as an asset. Moreover, the underlying index

16
Geltner and Fisher [2007] describe the major ideas related to equilibrium modeling in real estate.
17
Titman and Torous [1989], Buttimer, Kau, and Slawson [1997], Bjork and Clapham [2002] and Ciurlia
and Gheno [2008] provide models using the no-arbitrage framework with complete markets models. Otaka
and Kawaguchi [2002] take a step further and develop a model for incomplete markets that consists of a
security market where stocks, bonds, currencies, and derivative securities that are traded without friction, a
space market with the rents of buildings, and a property market where the prices of real properties are
determined.
18
See Patel and Pereira [2006] for a more detailed discussion.

17
is only an observable variable and not an asset that could be traded freely without any

frictions, although real estate portfolios mimicking exactly the index are a possibility for a

distinct class of investors. The new generation of products dependent on real-estate risks cannot

be priced in a risk-neutral framework either. This is due to the relationship between interest rate

risk along the term structure of interest rates and the risk triggers in RMBS space such as

defaults, prepayments, and arrears. Hence, a real-world measure approach based on statistical

calibration of historical data is required for pricing real-estate derivatives. A new method focused

on the market price of risk as a modeling tool has been proposed by Fabozzi and Tunaru [2009].

Their models take into consideration the mean-reversion to a nonlinear long-run trend of real-

estate indices.

Property derivatives markets are incomplete because the primary asset underpinning

this market suffers from lack of homogeneity. 19 However, similar properties in close

geographical proximity should have similar prices. This high correlation is helpful for cross-

hedging spot real-estate portfolios with futures on local indices. Nevertheless, the lack of

homogeneity implies that hedging with real-estate derivatives is always going to be less than

perfect. 20

An important characteristic to account for when modeling a real-estate underlying

index is reversion to a long-run trend. Similar to commodity markets where the underlying

asset is a consumption asset, the supply and demand forces on real-estate markets drive real-

estate prices in a different way than a stock financial index.

Let {X t }t ≥0 be a stochastic process representing a real-estate index. Then consider the

process on the log-scale Yt = ln( X t ) . In Exhibit 2, the monthly series of the Case-Shiller

Home Price Composite-10 Index is displayed on the log scale. There is a clear linear time

19
Black [1986] emphasizes that for the smooth functionality of derivatives, a homogeneous underlying
asset is helpful.
20
Real-estate traders managing portfolios with assets worth individually a large amount of money have
extra exposure to idiosyncratic risk that cannot be hedged away easily with property derivatives. See Baum
[1991].

18
trend for this data series. The same can be said for the IPD UK Monthly Index 21 since 1986

as illustrated in Exhibit 3 using a log-scale and for the Halifax House Price Index in Exhibit 4

based on a quarterly historical series. For all indices, the log-index series fluctuates around a

positive linear time trend and any model for pricing real-estate derivatives on these indices

should account for this statistical property.

INSERT EXHIBIT 2 HERE

INSERT EXHIBIT 3 HERE

INSERT EXHIBIT 4 HERE

Therefore, we can assume that {Yt }t ≥0 is a process that is mean reverting towards a

deterministic linear trend. In other words, the log-index is the sum of a zero-mean stationary

autoregressive Gaussian process and a deterministic linear trend. One could also fit nonlinear

trends if necessary. It is possible then to calculate the solution of this equation in closed form

which can be useful for pricing derivatives. Hence, a real-estate index can be modeled with a

geometric Ornstein-Uhlenbeck process that always ensures positive levels. Because the real-

estate market is incomplete, for pricing purposes one needs to specify exogenously the

market price of risk. For example, an equilibrium model could be used to determine the

forward price on an index and then the market price of risk inferred from there can be used to

price other derivatives contingent on the same index.

The same methodology can be applied using a two-factor model where the de-trended

log-index and a representative interest rate (Yt − α − βt , rt ) are jointly normally distributed.

The marginal distribution of Yt − α − β t is therefore Gaussian and consequently the

distribution of real-estate index X t is log-normal. The bivariate process (Yt − α − β t , rt ) has

21
The IPD Monthly Index is based on an open market appraised valuations of real buildings and it covered
3,700 properties worth around £47 billion in September 2006. It includes commercial and other investment
properties representing more than 90% of the combined value of the property assets held in UK unit trusts
and other unit linked property investment funds.

19
three econometric characteristics that make it suitable for modeling a real-estate index: (1) it

is trend stationary (i.e, the log-index may vary around a time trend); (2) the variance of index

returns does not grow indefinitely with time, and; (3) the model implies autocorrelations that

can be positive or negative. 22

It is critical to be able to model real-estate indices as close as possible to reality since

many mortgage-related securities are marked to model in the absence of a liquid market. A

large bias in forecasting future levels of a real-estate index will be reflected say, in marking

the profit and loss position of a real-estate position, and this could be extremely detrimental

to banks holding positions in these securities.

SUMMARY

In this paper, we review the role of property derivatives and discuss the instruments

available. Mortgage-backed securities require specialized interest hedging tools capable of

handling the embedded prepayment and default risk. Given the special characteristics

associated with the real-estate asset class, we emphasize the main points to be taken into

account when pricing property derivatives, this being a typical situation of incomplete

markets.

REFERENCES

Baum, Andrew. “Property Futures.” Journal of Property Valuation and Investment, 1 (1991),
pp. 235-240.

Black, Deborah. Success and Failure of Futures Contracts: Theory and Empirical Evidence.
Monograph Series in Finance and Economics, Monograph 1986-1. New York University,
1986.

22
Detailed calculations of autocorrelations are provided in Lo and Wang [1995]. The capability of a model
to reproduce a wide range of autocorrelations is important because then it is possible to match the empirical
findings in Fama and French [1988a, 1988b] and Lo and Mackinlay [1988], that equity portfolios are
positively autocorrelated at shorter horizons and negatively correlated at longer horizons.

20
Bjork, Tomas, and Eric Clapham. “On the Pricing of Real Estate Index Linked Swaps.”
Journal of Housing Economics, 11 (2002), pp. 418-432.

Buttimer, Richard J. Jr., James B. Kau, and Carlos V. Slawson. “A Model for Pricing Securities
Dependent upon a Real Estate Index.” Journal of Housing Economics, 6 (1997), pp. 16-30.

Carlton, Dennis W. “Futures Markets: Their Purpose, Their History, Their Growth, Their
Success and Failures.” Journal of Futures Markets, 4 (1984), pp. 237-271.

Case, Karl E., John M. Quigley, and Robert J. Shiller. “Home-Buyers, Housing, and the
Macroeconomy.” In Anthony Richards and Tim Robinson (Eds.), Asset Prices and Monetary
Policy, Reserve Bank of Australia, 2004, pp. 149-188.

Case, Karl E., and Robert J. Shiller. “The Efficiency of the Market for Single Family
Homes.” American Economic Review, 79 (1989), pp. 125-37.

Case Karl E. and Robert J. Shiller. “Forecasting Prices and Excess Returns in the Housing
Market.” AREUEA Journal 18 (1990), pp. 253-273.

Case, Karl E. and Rober J. Shiller. “Mortgage Default Risk and Real Estate Prices: The Use
of Index Based Futures and Options in Real Estate.” Journal of Housing Research, 7 (1996),
pp. 243-58.

Case, Karl E., Robert J. Shiller, and Allan N. Weiss. “Index-Based Futures and Options
Trading in Real Estate.” Journal of Portfolio Management, 19 (1993), pp. 83-92.

Ciurlia Pierangelo and Andrea Gheno. “A Model for Pricing Real Estate Derivatives with
Stochastic Interest Rates.” Munich Personal RePEc Archive, No. 9924 (2008).

Clayton, Jim. “Commercial Real Estate Derivatives: They’re Here… Well, Almost.” PREA
Quarterly, Winter (2007), pp. 68-71.

Englund, Peter, Min Hwang, and John Quigley. “Hedging Housing Risk.” Journal of Real
Estate Finance and Economics, 24 (2002), pp. 167-200.

Fabozzi, Frank and Radu Tunaru. “Pricing Models for Real-Estate Derivatives”, Working
Paper, Yale School of Management, 2009.

Fama, Eugene and Kenneth French. “Permanent and Temporary Components of Stock
Prices.” Journal of Political Economy, 96 (1988a), pp. 246-273.

Fama, Eugene and Kenneth French. “Dividend Yields and Expected Stock Returns.” Journal
of Financial Economics, 22 (1988b), pp. 3-26.

Fisher, Jeffrey D. “New Strategies for Commercial Real Estate Investment and Risk
Management.” Journal of Portfolio Management, 32 (2005), pp. 154-161.

21
Horrigan, Holly T., Bradford Case, David Geltner, and Henry O. Pollakowski. “REIT-Based
Commercial Property Return Indices: A Model to Support and Improve Segment-Specific
Investment in the Real Estate Markets.” MIT Working Paper, 2009.

Geltner, David and Jeffrey Fisher. “Pricing and Index Considerations in Commercial Real
Estate Derivatives.” Journal of Portfolio Management, Special Real Estate Issue (2007), pp.
99 -117.

Lo, Andrew and Craig MacKinlay. “Stock Market Prices Do Not Follow Random Walks:
Evidence from a Simple Specification Test.” Review of Financial Studies, 1 (1988), pp. 41-
66.

Lo, Andrew and Jian Wang. “Implementing Option Pricing Models When Asset Returns Are
Predictable.” Journal of Finance, 50 (2005), pp. 87-129.

Otaka, Masaaki and Yuichiro Kawaguchi. “Hedging and Pricing of Real Estate Securities
under Market Incompleteness.” MTB Investment Technology Institute Co., Ltd., Tokyo
(2002).

Patel, Kanak and Ricardo Pereira. “Pricing Property Index Linked Swaps with Counterparty
Default Risk.” Journal of Real Estate Finance and Economics 36 (2008), pp. 5-21.

Pavlov, Andrey D. “Competing Risks of Mortgage Termination: Who Refinances, Who

Moves, and Who Defaults?” Journal of Real Estate Finance and Economics, 23 (2001), pp.
185-211.

Seiler Michael J., James R. Webb, and F.C. Neil Myer. “Diversification Issues in Real Estate
Investment.” Journal of Real Estate Literature, 7 (1999), pp. 163-179.

Shiller Robert J. and Allan N. Weiss. “Home Equity Insurance.” Journal of Real Estate
Finance and Economics, 19 (1999), pp. 21-47.

Shiller, Robert J. “Derivatives Markets for Home Prices.” Yale Economics Department
Working Paper No. 46, Cowles Foundation Discussion Paper No. 1648 (2008).

Shiller, Robert J. “Measuring Asset Value for Cash Settlement in Derivative Markets:
Hedonic Repeated Measures Indices and Perpetual Futures.” Journal of Finance 68 (1993),
pp. 911-931.

Stulz, René. “Should We Fear Derivatives?” Journal of Economic Perspectives, 18 (2004),

pp. 173-192.

Titman, Sheridan and Walter Torous. “Valuing Commercial Mortgages: An Empirical

Investigation of the Contingent-Claim Approach to Pricing Risky Debt.” Journal of Finance,
44 (1989), pp. 345–373.

Webb, James R., R.J. Curcio, and Jack H. Rubens. “Diversification Gains from Including
Real Estate in Mixed-Asset Portfolios.” Decision Sciences, 19 (1988), pp. 434-452.

22
 Pool of mortgages

SVR Fixed s
coupons coupons upon
SV R co Interest rate swap desk
d or
Fixe
ad
spre
R+
Bank owning loans L IB O

Portf
o lio of
in
Libor + spread deriv terest ra Interest rate derivatives desk
(swa ative te
ption s
s, ca
ps, f
loors
)

Investors

Exhibit 1. The main components affecting a balance guaranteed swap linked to a

portfolio of mortgage loans. The SVR is the standard variable rate that banks charge their
mortgage borrowers when their mortgage switches from fixed the floating. It is driven by
the official base rate but its exact value is at the discretion of the bank.

23
 Case-Shiller Home Price Composite 10 Index on Log
Scale

5.5

5
Log CSI

4.5

3.5

3
Jan-87

Jan-89

Jan-91

Jan-93

Jan-95

Jan-97

Jan-99

Jan-01

Jan-03

Jan-05

Jan-07
Exhibit 2. Historical monthly data of Case-Shiller Home Price Composite 10 Index,
monthly on the log scale. The series displays a linear time trend.

24
 IPD UK All Property Index on Log scale

7.5

6.5

5.5
Index

4.5

3.5

3
Dec-86

Dec-87

Dec-88

Dec-89

Dec-90

Dec-91

Dec-92

Dec-93

Dec-94

Dec-95

Dec-96

Dec-97

Dec-98
Dec-99

Dec-00
Dec-01

Dec-02

Dec-03

Dec-04

Dec-05

Dec-06

Dec-07

Exhibit 3. Historical monthly data of IPD UK Monthly Index on the log scale. The series
displays a linear time trend.

25
 Halifax House Price Index on Log scale

6.5

5.5
Index

4.5

3.5

3
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
Exhibit 4. Historical quarterly data of Halifax House Price Index (HHPI) on the log
scale.

26